Project overview
ConnectAI's call-center solution required efficient post-call processing — saving recordings and analyzing call data. The previous event-driven design with SNS/Lambda chains was complex, hard to debug and prone to data inconsistency.
Challenges
- Complex chaining of SNS, Lambda and S3 events
- Difficult root-cause analysis across distributed events
- Async ordering caused missed or out-of-order processing
- Limited observability across services
- High operational overhead
Our approach
We replaced the event-chain architecture with AWS Step Functions to orchestrate the entire post-call workflow.
AWS Step Functions orchestration
- State-machine orchestration from fetching recordings to analysis and storage
- Sequential and parallel execution paths for optimal data flow
- Built-in error handling, retries and failure paths for fault tolerance
- Visual workflow with real-time execution states, logs and metrics
Outcomes
- Errors reduced by 90% with consistent workflows
- Root-cause analysis time cut by 70%
- 100% consistency in post-call data processing
- Operational overhead reduced by 50%